|
Home Page
Demos
Try ILOG Discovery
Use it in your
application
FAQ
Contact
|
Overviews give you a first overall glimpse of your data set when the data consists of numerous numerical or enumerated variables. These visualizations let you immediately spot the general distribution of each variable, find out if you have some outliers or missing data, and quickly detect possible correlations between variables. It is a good idea to start your data exploration with one of these views:
-
Parallel coordinates.
-
Scatter plot matrices.
-
Distributed histograms.
-
A variety of other overviews can be manually created.
 |
Parallel Coordinates Let you view all numerical and enumerated columns of a
data table side by side. The value held for each object is represented
by a polygonal line. This lets you find easily which columns are correlated with
one another, and the similarities in distribution.
|
 |
Scatter plot matrix
Presents you a matrix of correlation of each column against each other.
This view is very powerful, as in one glance, you can spot all
correlation. It is however not suited to large data sets. |
 |
Distributed histograms.
Given a main variable (here, the longitude of a set of cities),
all objects are distributed along the horizontal or vertical axis
according to that variable's value. local histograms or bar codes for
each individual object give an idea of how other attributes are related
to the main axis, or with one another.
|
|
